Test of Beta 2a.

For topics about current BETA or future releases, including feature requests.

Test of Beta 2a.

Postby MikeGale » Mon Sep 18, 2006 6:44 pm

Hi,

I've just run through a few tests of Beta 2a.

The tests comprised about twenty pages, containing, HTML 4 , XHTML 1 transitional, some encoded as ANSI, some as utf-8, windows, unix, mac line ends, asp code embedded and aspx code embedded.

These are validation tests, with my custom setup and both with and without nsgmls.

All run by hand and brief notes made.

Results are:

1) Some odd things happened when I installed with IE running. Long story short, several attempts to re-install did not get it working from other applications, so I refreshed the version 7 installation (in parallel) to get things working again.

2) The utf-8 encoded pages did not get evaluated by nsgmls. The byte order marks at the beginning of the file threw nsgmls out of kilter. Saving the same page as ANSI worked as expected.

3) A line with a quotation removed ("<script src="/_ScriptLibrary/UtilitiesDE.js type="text/javascript" language="Jav...") continued to throw 79 errors (in CSE) but nsgml detected the issue. This suggests that there might be a way of leveraging nsgmls to improve the detection of CSE in this case.

4) Running checks on both ASP and ASP.NET raw source (both with code on the markup page) with nsgmls was (unsurprisingly) not very good. Without the nsgmls it was better than expected.

5) In one case code inserted content ("<b'anglebracket here'<%= ...></b") was interpreted as no content within the markup. In other words it gave an error. If this had been interpreted as a non-empty string it would have been better.

6) I noticed one place in the help file that suggested that the Tidy and nsgmls were on the same tab.
User avatar
MikeGale
Rank VI - Professional
Rank VI - Professional
 
Posts: 612
Joined: Mon Dec 13, 2004 2:50 pm
Location: Tannhauser Gate

Postby Albert Wiersch » Tue Sep 19, 2006 12:14 pm

Thanks Mike!

1) I don't know why IE should interfere with installation. I've never had this issue. The issue that could affect CSE HTML Validator with other applications is if CSE HTML Validator is not installed with an account with administrator privileges.

2) It will be interesting to see how many people will use the nsgmls functionality. If a lot, then I may look to see if there is anything I can do about this issue with UTF-8, or perhaps switch to a different validator that can handle it in the future.

3) I still hope to get better missing quote handling in v8.0x, perhaps in v8.01 or 8.02.

4) I'm glad to hear that. I did some work to make validating raw ASP code better.

5) Is this a big issue for you? Do you think this is common enough to warrant addressing? I may be able to address this if so.

6) I believe I found the place where it implied it (and fixed it). Originally I was going to use one tab for both HTML Tidy and nsgmls but decided to separate each into their own tab.
Image
Albert Wiersch
User avatar
Albert Wiersch
Site Admin
Site Admin
 
Posts: 2435
Joined: Sat Dec 11, 2004 10:23 am
Location: Near Dallas, TX

Postby MikeGale » Tue Sep 19, 2006 3:08 pm

Some quick feedback:

1) I had a non-standard installation. 8 (Beta 1) installed but 7 running from the edit button of IE etc. I installed with IE still active. I'm not greatly concerned about this.

2) It occurred to me that it may be easy to create a version of the source that is ANSI encoded and submit that to nsgmls. (I did something akin to this when I transformed content in .NET before submitting it to the DLL, to get line numbering in synch.)

5) That %= style of in-line replacement is not something I do often. So it's not serious for me. I have seen code samples that use it heavily, so some CSE users might benefit.
User avatar
MikeGale
Rank VI - Professional
Rank VI - Professional
 
Posts: 612
Joined: Mon Dec 13, 2004 2:50 pm
Location: Tannhauser Gate

Postby Albert Wiersch » Wed Sep 27, 2006 12:10 pm

To follow up on #3, I was able to add "missing quote detection" in BETA 3 (not yet available). This is brand new so it hasn't been thoroughly tested yet, but it can make a huge difference in the results when there are missing quotation marks. I hope to have a final BETA out at the end of this week so this can be better tested by users.

I'm glad I finally got something like this in, as missing quotation marks have been something CSE HTML Validator did not handle very well in the past.

The main issue with this new feature is detecting missing quotation marks when there really aren't any. I hope this rarely, if ever, happens, but it is a possibility. I'd love to get feedback on this when BETA 3 is released.

If there are any problems with this new feature, it can be turned off in the Validator Engine Options and CSE HTML Validator will work like before.
Image
Albert Wiersch
User avatar
Albert Wiersch
Site Admin
Site Admin
 
Posts: 2435
Joined: Sat Dec 11, 2004 10:23 am
Location: Near Dallas, TX

Postby MikeGale » Wed Sep 27, 2006 2:13 pm

That's very good.

Looking forward to it.
User avatar
MikeGale
Rank VI - Professional
Rank VI - Professional
 
Posts: 612
Joined: Mon Dec 13, 2004 2:50 pm
Location: Tannhauser Gate


Return to CSE BETA Talk

Who is online

Users browsing this forum: No registered users and 2 guests